[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#953292: RFS: c-icap/1:0.5.6-1 [QA] -- ICAP server implementation



On Mon, Mar 09, 2020 at 04:37:17PM +0100, Mattia Rizzolo wrote:
> On Sun, Mar 08, 2020 at 01:00:34AM +0100, Adam Borowski wrote:
> > On Sat, Mar 07, 2020 at 10:26:05AM +0100, Håvard Flaget Aasen wrote:
> > >    * Change path from /var/run/ to /run/ in postinst and postrm scripts
> > 
> > Hi!
> > I don't understand the /var/run/ -> /run/ change -- it looks grossly
> > incomplete.  You change it only for the newly created user, but:
> > * systems which ever had a previous version of c-icap will have the
> >   user's home in /var/run/c-icap
> 
> Incidentally, ssh itself did this, and indeed in my system:
> 
> sshd:x:107:65534::/var/run/sshd:/usr/sbin/nologin
> sshd:x:101:65534::/run/sshd:/usr/sbin/nologin

> After seeing that ssh didn't really care about transitioning between the
> directories I also sponsored another package changing directories
> without bothering about it.

Well yeah -- the home dir for ssh is not even used -- sshd.pid is managed by
start-stop-daemon (no idea about a certain racy init/rc), with /run/sshd
being an empty directory (that nevertheless gets recreated at start).

For c-icap, though, it's used from a number of places, so I was a little
scared about relying that every machine has the symlink.

> > * postrm will remove only the new location (on most systems /var/run/ is a
> >   symlink to /run/, though)
> 
> Consider that /run is a tmpfs in a standard system, so this is not
> *that* important.

Yeah but /var/run might not be, if it's not a symlink.  One can argue it
should always be, but...


Meow!
-- 
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ in the beginning was the boot and root floppies and they were good.
⢿⡄⠘⠷⠚⠋⠀                                       -- <willmore> on #linux-sunxi
⠈⠳⣄⠀⠀⠀⠀


Reply to: